Benchmarks

FastQ vs The World

Real numbers. Same hardware. No cheating.

Tested on: Debian 13 · Intel Core i7 (13th Gen) · 64 GB RAM · Redis 8.0 local · consumer laptop

27k/s
FastQ push throughput
5k/s
BullMQ throughput
8 MB
FastQ memory
100 MB
BullMQ memory

Throughput (jobs/sec, higher = better)

FastQ 27,000 jobs/s
Sidekiq 7,200 jobs/s
BullMQ 5,100 jobs/s
Celery 3,900 jobs/s

Memory usage (MB, lower = better)

FastQ 8 MB
Sidekiq 80 MB
BullMQ 100 MB
Celery 120 MB

Live dashboard

Live dashboard coming soon.

Methodology

Test environment

  • • Debian 13
  • • Intel Core i7 (13th Gen)
  • • 64 GB RAM
  • • NVIDIA GeForce RTX 4060
  • • Redis 8.0 (local, no network latency)
  • • Consumer laptop (not server hardware)

What was measured

  • • 10,000 jobs (200 warmup) per run
  • • Peak throughput (jobs/sec, push & pop separately)
  • • Idle RSS memory after startup
  • • No GC tuning, default config for all
  • • p50 and p99 latencies recorded

Reproduce these benchmarks: github.com/OxoGhost01/FastQ/benchmarks

A note on competitor numbers: Some frameworks (e.g. BullMQ) publish throughput figures measured on dedicated server hardware under ideal conditions — numbers like 250 k/s are achievable on high-core-count machines with NVMe storage and tuned Redis. FastQ's benchmarks above were intentionally run on a consumer laptop to provide a conservative, reproducible baseline. On equivalent hardware, FastQ consistently outperforms BullMQ, Sidekiq, and Celery because it has no managed runtime overhead — just C, Redis, and pthreads.